GABI Genomic Analysis of Bacterial Isolates

Descriptive metrics of individual assemblies determined by Quast.

Insert size refers to the size of the sequenced DNA fragment. Depending on the exact library protocol, this size will fall fairly uniformly around a mean value (~300-500bp). For Illumina data, that value should typically be (slightly) larger than the combined length of forward and reverse read for optimal data yield. Very flat curves may (depending on the protocol!) indicate a failure during fragment size selection/enrichment. Neither small insert sizes nor flat curves are a clear predictor for subsequent assembly issues, but can inform any potential debugging efforts.

BUSCO scores describe the coverage of the assemblied gene space against a set of broadly conserved singleton genes (here: bacteria_odb10). A perfect assembly should have a complete coverage of the gene space (complete: 100%), without any fragmentation or, worse, duplication. A high value of duplication may indicate assembly errors or contamination. Some taxa with very streamlined gene content, such as Campylobacter, will typically have a completeness score of less than 100%. The Completeness estimates may include duplicated genes, so values greater than 100% are possible (i.e. all genes present, of which x % are duplicated).

Bracken processes raw outputs from Kraken2, which matches kmers from raw sequencing reads against a reference database to determine the taxonomic composition of a read set. For DNA from pure cultures (which is the focus of GABI), only one species should be identified at dominant proportions. For some taxa, like Campylobacter, several species from the same genus may be found at comparative abundances due to a lack of sufficient DNA differences. Otherwise, identification of multiple taxa at higher proportions may indicate a contamination issue.

Taxa-specific MLST schemas classify assemblies into pre-defined types or groups. Results are divided by typing schema (and consequently taxa).

Serotyes, similar to MLST types, classify assemblies based on a set of predefined gene profiles.